INF2404H — Special Topics in Information: Explainability & Fairness for Responsible Machine Learning
Machine Learning applications are increasingly utilized to make crucial decisions in many sectors of our economy and society. These include, but are not limited to, healthcare, financial services, public safety, and higher education. Predictions from machine learning systems are incorporated within organizational processes to support evidence-based decision-making. This course examines state of the art techniques and technologies related to explainability and fairness in machine learning applications. These human-centric aspects play a significant role in the design and operation of machine learning applications. Absence of explainability and fairness capabilities in a machine learning application erodes its public legitimacy and undermines its social licence. This reduces its acceptance and adoption in the real-world. Students will use frameworks and techniques for architectural modeling, analysis, and design to understand explainability and fairness in the context of machine learning applications.